Every Community is a Startup
CLS Session E, 4:30pm Every Community is a Startup Organizer: Kevin Shock @shockeyk Notes: Jeffrey Sens, @jeffreysens Attendees: Patrick Cozzi, John Fitzpatrick, Richard Meinsen, Heather Leson, Jeffrey Sens, Kaliya Hamlin (Identity Woman) What is the image of the lean startup? Wile E Coyote? The assumptions should not be taken for granted and should be checked as quickly/efficiently as possible. Build, Measure & Learn. Need to capture metrics based on real behavior--dangers of surveys and research, and create a test for the MVP (minimum viable product). For community managers, how can we think about community organizing as MVP: vision is a basic test of MVP moment. You might think the vision is beautiful, but testing the vision is essential. Mistaking likes for other kinds of behavior. How synced up are you? If one part of the development team is in a bubble, they may not understand users at all. One case, how much a software developer changed their work when they actually went to a community event, which creates accountability. For commercial products, you can say there is a community of users and focus groups can give you lots of perceived information (I want X, I want Y). But these self-reports are themselves abstract from the actual behavior. We have many different kinds of community managers, and the best events of actual feedback are where people have something they really want to know about. A/B Testing with Google Ads to measure the differential between desired and undesired action. Confouds--multiple things may produce the effect you’re looking for but you don’t know which. How do norms of our community impact testing? For example, many A/B tests rely on people not knowing they are involved in a test, that we are data mining their choices. People are not black boxes and ethics/consent do matter. Note: this is a lot like sensitivity analysis, an economic projection model where we change an assumption by X, how much does that change Y? We learn most not from succeeding or failing but succeeding after failing. How do we test for engagement and taking action, for example, filing a bug report? Divide up into two e-mail sets, and vary the two approaches, e.g. one sent out end of day and other at beginning of day. What are the limits of testing? The diminishing returns of constant testing? Is this mainly about being aware of assumptions? Growing community can take inspiration from MVP, to make informed choices about where to focus energy and modify. High Performance Collaboration wiki, Eugene Kim, helped us to do this. Pattern Language for groups and meeting, developed by multi disciplinary facilitators. Many only were expert in one method and locale (e.g. consensus) but they practices multiple patterns. MVP is about what sticks and iterating. Finding a sub-community within a larger community. We found each other, create a mailing list and blogging in early 2005. False starts in hosting our own conference, but we kept meeting people at other recurring conferences to build momentum. Community Testing Methodology--developed by a group OSIS which may model Lean Startup for community building. http://osis.idcommons.net/wiki/Main_Page